Target tracking using adaptive compressive sensing and processing

• Adaptive compressive sensing and processing is applied to a single target tracking scenario.
• A particle filtering method based on likelihood sampling is proposed to sequentially estimate the target posterior.
• A Monte Carlo based method is provided that predicts the probability of appearance of signal elements in the measurements one time step ahead in the tracking scenario.
• The adaptive sensing matrix is naturally configured for each radar sensor independently based on sequentially updated information on the target state within the target tracking scenario.

Compressive sensing and processing (CSP) performs signal acquisition and processing at sub-Nyquist rates. This makes CSP an attractive option in radar target tracking as it reduces computational load in sequential signal acquisition and processing. However, CSP is accompanied by a reduction in signal to noise ratio (SNR) which results to a deterioration of tracking performance. In order to improve tracking performance CSP can be configured using information available on the target state which is provided by the sequential estimation process. In this work, adaptive CSP is applied to a target tracking scenario. A particle filtering implementation using adaptive CSP is developed for tracking a single target. Simulation results are provided to demonstrate the improvement in SNR and in tracking a single target by adaptive CSP over non-adaptive CSP and previously proposed adaptive CSP methods.